Papers › Frozen Transformers in Language Models Are Effective Visual Encoder Layers

Frozen Transformers in Language Models Are Effective Visual Encoder Layers

19 Oct 2023arXiv:2310.12973archive 2025-07-28

Ziqi Pang, Ziyang Xie, Yunze Man, Yu-Xiong Wang

This paper reveals that large language models (LLMs), despite being trained solely on textual data, are surprisingly strong encoders for purely visual tasks in the absence of language. Even more intriguingly, this can be achieved by a simple yet previously overlooked strategy -- employing a frozen transformer block from pre-trained LLMs as a constituent encoder layer to directly process visual tokens. Our work pushes the boundaries of leveraging LLMs for computer vision tasks, significantly departing from conventional practices that typically necessitate a multi-modal vision-language setup with associated language prompts, inputs, or outputs. We demonstrate that our approach consistently enhances performance across a diverse range of tasks, encompassing pure 2D and 3D visual recognition tasks (e.g., image and point cloud classification), temporal modeling tasks (e.g., action recognition), non-semantic tasks (e.g., motion forecasting), and multi-modal tasks (e.g., 2D/3D visual question answering and image-text retrieval). Such improvements are a general phenomenon, applicable to various types of LLMs (e.g., LLaMA and OPT) and different LLM transformer blocks. We additionally propose the information filtering hypothesis to explain the effectiveness of pre-trained LLMs in visual encoding -- the pre-trained LLM transformer blocks discern informative visual tokens and further amplify their effect. This hypothesis is empirically supported by the observation that the feature activation, after training with LLM transformer blocks, exhibits a stronger focus on relevant regions. We hope that our work inspires new perspectives on utilizing LLMs and deepening our understanding of their underlying mechanisms. Code is available at https://github.com/ziqipang/LM4VisualEncoding.

PaperPDFCodeCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 16 code samples harvested from 1 repository linked to this paper: 12 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 3 honoured, 1 violated, 3 with no contract checked; 5 where Syntology's instrument failed) · 4 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; 3 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 16 samples from 1 repository, 12 ran (7 with no instrument failure; 5 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 12 samples that ran has no recorded date.

ziqipang/lm4visualencoding officialmentioned in papermentioned on GitHubpytorchMITSyntology: 16 samples, 12 ran (5 instrument-failed) · 4 unverified report
zhixinlai/llmboostmedical mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Residual-based Language Models are Free Boosters for Biomedical Imaging (arXiv:2403.17343; that paper's own run record: 12 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 1 honoured, 3 violated, 2 with no contract checked; 6 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method. report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

16 samples harvested: 12 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 3 honoured, 1 violated, 3 with no contract checked; 5 where Syntology's instrument failed) · 4 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

3ran · honoured contract
1ran · violated contract
2ran · our draft was wrong
3ran; the contract check could not run
3ran
4unverified

Licence: 8 of the 16 samples are pointer only, meaning Syntology does not serve that copy's text. For 8 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from ziqipang/lm4visualencoding. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

ziqipang/lm4visualencoding official repository 16 samples, 12 ran (5 instrument-failed) · 4 unverified
add_weight_decay in 27 papers image_classification/optimizer_utils.py ran MIT (permissive) · e805e183bad0f027 · report fixture: signature model:callable/0/?, weight_decay:float/0/float, skip_list:tuple/1/str, param_filter:NoneType/0/? · returned list · level recorded 2026-09-24
apply_rotary_emb in 121 papers image_classification/models/llama.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant MIT (permissive) here; the stored copy is pointer only · b47d48e431b34acd · report fixture: signature xq:Tensor/4/float, xk:Tensor/4/float, freqs_cis:Tensor/2/complex · returned tuple · level recorded 2026-09-22
build_transform in 170 papers image_classification/datasets.py ran · our draft was wrong MIT (permissive) · 3bc74137ab36aa79 · report fixture: signature is_train:bool/0/bool, args:ArgsConfig/0/? · returned Compose · level recorded 2026-09-26
get_loss_scale_for_deepspeed in 20 papers video_understanding/engine_for_finetuning.py ran · honoured contract MIT (permissive) here; the stored copy is pointer only · 29416043c7035e7c · report fixture: signature model:callable/0/? · returned float · level recorded 2026-09-25
get_mce_from_accuracy in 5 papers image_classification/utils.py ran · honoured contract fingerprinted MIT (permissive) · a9e48df75d2e823e · report fixture: signature accuracy:float/0/float, error_alexnet:float/0/float · returned float · level recorded 2026-09-24
get_sinusoid_encoding_table in 51 papers video_understanding/modeling_finetune.py ran · honoured contract fingerprinted MIT (permissive) here; the stored copy is pointer only · da651e3979a18f84 · report fixture: signature n_position:int/0/int, d_hid:int/0/int · returned Tensor · level recorded 2026-09-22
optimizer_kwargs in 13 papers image_classification/optimizer_utils.py ran · our draft was wrong MIT (permissive) here; the stored copy is pointer only · 2cd1e282432041c8 · report fixture: signature cfg:SimpleNamespace/0/? · returned dict · level recorded 2026-09-24
precompute_freqs_cis in 112 papers image_classification/models/llama.py ran · violated contract MIT (permissive) here; the stored copy is pointer only · 14a84c2cbfebc413 · report fixture: signature dim:int/0/int, end:int/0/int, theta:float/0/float · returned Tensor · level recorded 2026-09-22
reshape_for_broadcast in 90 papers
census: 6 implementations from 10 papers, 1 distinct output on one shared input
image_classification/models/llama.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted MIT (permissive) here; the stored copy is pointer only · 70bf6ebaafd266c4 · report fixture: signature freqs_cis:Tensor/2/float, x:Tensor/4/float · returned Tensor · level recorded, date not recorded
tensor_normalize in 18 papers video_understanding/kinetics.py ran MIT (permissive) here; the stored copy is pointer only · 14b3b13b2e7ee844 · report fixture: signature tensor:Tensor/2/int, mean:list/1/float, std:list/1/float · returned Tensor · level recorded 2026-09-25
train_class_batch in 17 papers video_understanding/engine_for_finetuning.py ran MIT (permissive) here; the stored copy is pointer only · c7976ea27edc377a · report fixture: signature model:callable/0/?, samples:Tensor/2/float, target:Tensor/1/int, criterion:callable/0/? · returned tuple · level recorded 2026-09-25
unnormalize_image in 5 papers image_classification/utils.py ran; the contract check could not run (RAISES) fingerprinted MIT (permissive) · 32208d3133397a32 · report fixture: signature data:Tensor/4/float · returned ndarray · level recorded 2026-09-24
PGDAttack in 5 papers image_classification/engine.py unverified MIT (permissive) · 58ab30e8c22165b6 · report
build_dataset in 151 papers image_classification/datasets.py unverified MIT (permissive) · 8b43480472330eaf · report
patch_level_aug image_classification/engine.py unverified MIT (permissive) · a57f32dae793402d · report

Tasks

Action RecognitionImage-text RetrievalMotion ForecastingPoint Cloud ClassificationQuestion AnsweringText RetrievalVisual Question Answering

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Question Answering SQA3D LM4VisualEncoding AnswerExactMatch (Question Answering) 48.09 #4 of 7 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Focus

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections